a type of structural equation modeling used to examine a set of simultaneous linear relationships between variables. Results are displayed using a figure in which boxes denote measured variables, bidirectional arrows show correlations, directional arrows show causal or predictive relations, and arrows coming from nowhere show error associated with the prediction model. Each variable is expressed as a linear function of the preceding variable as well as by a unique latent variable; the weights and standard errors for each coefficient provide information about the size of the parameter. Goodness of fit is tested by comparing models with more pathways to models with fewer pathways. Also called path modeling. See path diagram.